Correction: in the Normal cycle, it's 2 full revolutions in one direction, then a one-second pause, then 1-1/2 revolutions in the other direction. Sounds like there's a similarity with Cybrvanr's Whirlpool compact.
Retromom:
These things are apparently still made in huge numbers in Asia (see search instructions below) and are also popular in the dry regions of Australia.
They're made in various sizes from compact (5 lb.) to large (16 lb.), so size isn't the key point. Only two are currently imported into the US: Danby is rated max 10 lbs. (I would say 8 to be safe and so far I've run loads at about 6 lbs. to get a sense of how it performs), and a new Haier is rated at 16 (I would say 12 to be safe) (see, there's a pattern here: I tend to treat my tools carefully in general, washers included).
Electricity efficiency: Power consumption with all three motors running (washtub, spintub, pump), is 500 watts while doing 8 lbs. My TL Hotpoint on the other hand (Spiralator-equipped if I've got my terminology correct) uses 930 watts while doing 10 lbs. (Since hanging out around here I realized it's OK to have more than one!

Assuming equal cycle length, the Danby is using 62.5 watts per lb., and the Hotpoint is using 93 watts / lb. So the Danby is about 33% more efficient, though they are both useful to have.
I'm going to get more specific and empirical figures for actual kilowatt-hours per load as I run my experiments, and will post those when I have them.
Water efficiency: The "final rinse" water is basically clean water with a little bit of residual detergent, so it can be kept in the washtub for use as washwater for the following cycle. That's between a 30% and 50% efficiency gain, depending on how one handles the rinsing (which in turn depends on quantity of detergent in the wash cycle, thus more experiments to run).
And all this for $250 plus shipping, compared to minimum twice that for a decent (Danby, again!) front-loader. Which is the next major benefit: savings on first cost, and therefore more accessible to apartment-dwellers.
And as PeterH points out, simple mechanism, which translates to a theoretically lower potential for something to get out of order, and also less expensive repairs.
Niche market: Depends. Niche market in the USA presently. Major market in the UK in the 50s and 60s, and presently in Asia, Australia, and the Pacific Islands. I suspect market share in North America and elsewhere will grow as water & energy issues become more critical. People on a tighter budget will view these things as a way to gain efficiencies at lower cost. $250 for a twin tub, vs. minimum $500 for a front loader, will make a significant difference to many potential buyers.
Also potentially faster doing the laundry altogether. You have one load spinning while another is washing. Theoretically, running cycle times in parallel like that is twice as fast as any machine using comparable cycle times in series; though in a practical sense it's not quite, because you spend a minute or two transferring clothes during each cycle change. Also there's a learning curve, as with anything requiring manual input, it will take a while to get "good at it" to the point where I can do the whole routine using "muscle memory" without having to think much about it. At which point, doing the laundry becomes a good time to have a long phone conversation. Except it's kinda' fascinating to watch, so...

.
And, as Foraloysius pointed out, fun to play with!

. Seriously though, one of my main reasons for buying this unit is my R&D project on graywater recycling: I need a machine that's in current production and has manual controls, in order to have control over all of the variables: wash time, rinse time, spin time, water input / quantity / discharge, number of rinse cycles, etc.
I think the ideal machine for that purpose would be a twin tub with a horizontal drum, loaded through a hatch at the top (and with glass at the front of the drum that lined up with a round window on the cabinet, so you can see the water & clothing levels and watch the operation). And manual controls for speed of rotation and frequency of rotation-reversals. I don't know that anyone's done such a thing yet, so I guess that's my first contribution to the "invent a new washer" discussion (which I will be getting back to over the weekend, finally!).
Here's how to get a glimpse of how popular these things are in certain parts of the world:
Go to
www.alltheweb.com, click the search for Pictures, and put in the phrase "twin tub" as two words (with a space in the middle) and without the quote marks around it. Click the Search button and look at what comes up. All of those are made in Asia, mostly in China. To that add the other known manufacturers, some of which are big names: Panasonic, Toshiba, LG, Danby, Haier, and one or two more if I'm not mistaken. Whirlpool even makes one over there, but doesn't import it into the USA. Yet.
So you can see, this is a part of the industry that hasn't gotten a lot of attention in the USA yet, but it's by no means small. In the near future we might expect some higher-priced models with features such as stainless steel tubs and built-in water heaters, spin-rinse functions, and auto-soak functions such as Panasonic's Soakmatic feature.